Following is a short introduction to measurement of some types of electrical brain activity associated with motor control and what is currently believed to be their meaning. It should be noted that application of the invention is not necessarily constrained by these meanings and other signals may be measured and/or the following signals be used in other ways. Various articles are listed at the end of the background section.
Movement-Associated Cortical Potential (MAC) Accompanying Voluntary Movement
Experiments have shown that every voluntary movement is associated with an electrical cortical potential that can be recorded over the scalp. This activity is typically characterized by three components:                1. The “Bereitschaftspotential” (BP) or “Readiness Potential” defined as a slowly ‘rising’ negative potential that occurs 1-2 seconds prior to volitional self-initiated movements. It is related to the preparatory process prior to limb movement.        This BP consists in fact of two components:                    an early component (BP1) that lasts from the very beginning of the BP (starting 1-2 s or more prior to movement onset depending on the complexity of the movement) to approximately 0.5 s before movement onset; and            a late component (BP2) that occurs for the last half second before onset (see FIG. 1). BP2 has a steeper negative slope than BP1.                        2. The motor potential (MP) which consists of an initial sharp negative deflection that follows the BP's more gradual negativity. This potential is related to motor activity. At movement onset (at t=0 as shown in FIG. 1 below), there exists a sharp positive inflection that peaks at around 200 ms after the movement onset. This period is typically contaminated with EMG artifacts.        3. The post-movement activity (PMA) which is the potential change (starting more than 200 ms after the movement onset) whereby the brain resynchronizes and resumes ‘normal’ activity.        
FIG. 1 presents an averaged Motor Related Potential (MRP) template that illustrates these distinguishable periods. This in an example of an averaged MRP recorded for 918 left finger movement trials (onset at t=0) at C3 (channel 3) and C4 (channel 4).
For unilateral movement, BP1 typically has a symmetric and bilateral topography on the scalp, i.e. it is not lateralized about the motor cortex.
In contrast BP2 is typically larger (more negative) over the primary motor area of the contra lateral hemisphere. This is evident in FIG. 1 for the last ˜200 ms prior to finger press at time t=0. The electrode C4 is positioned on the right side of the head and for the left finger movement as shown exhibits a more negative potential on average than the contra-laterally placed C3 electrode.
Rich experimental evidence indicates that BP1 and BP2 might involve different functional systems. Experiments in PET (Positron Emission Tomography) and unicellular recordings in monkeys suggested that parts of the mesial frontal cortex, and typically the Supplemental Motor Area, may be involved in the generation of BP1. On the other hand, several investigators concluded that BP2 potential reflects expression of nerve excitation, namely, activity of cortical-spinal tract concerning efferent discharges of pyramidal tract.
It has been suggested that the awareness of willingness to move occurred later than the beginning of the electrophysiological event and that, consequently, the first part of the decision process to move was infra-conscious, at least for self-paced tasks.
All the above supports the fundamental EEG theory that potential negativity can be related to activity of the cortical areas whereas positivity is related to inactivity. Since the extremities of the human body are controlled by the contra lateral side of the brain it is generally expected that there should be more activity and hence ‘negativity’ on the contra lateral side. However, it should be noted that this is not always the case.
It has been shown that, the signal distribution over the scalp of the BP2 potential shows maximum at C3, (central left scalp), in case of voluntary right upper arm flexion movement. The maximum was at C4, (central right scalp), in case of voluntary left upper arm flexion movement. The distribution of the late PMA potential showed maximum at Cz, (central medial) in case of voluntary right or left upper arm flexion movement. The only part of the MAC that shows potentials contra lateral to the side of the movement is BP2.
Contingent Negative Variation (CNV)
In some experimental setups the generation of a MAC potential involves the performance of a prescribed task under the prompting of a pair of cuing stimuli: S1 and S2, separated by a given time interval. The first cue, (S1), is a ‘warning’ or ‘preparatory’ cue which is subsequently followed by a second ‘imperative’ cue (S2).
The subject is instructed to perform the given task as fast as possible following the presentation of the imperative stimulus (S2). Briefly, the preparatory stimulus precedes the imperative and thus acts as a ‘get ready’ signal to warn the subject that the imperative stimulus is approaching.
Under these conditions, the resultant waveform recorded over the scalp is a slow negative shift beginning at the presentation of S1 and ending roughly at the presentation of S2.
FIG. 2 shows a typical event-related potential in an S1-S2 paradigm, measured from central derivations (average of C3 and C4). The x-axis shows the presentations of S1 (onset at time t=0), and S2 (onset at t=6 seconds). The measures for PSW, NSW, and CNV (shown on the figure) as used in the present study are indicated.
The task can be, for instance, that S1 is a sound which the subject had to decide if it belongs to a previously memorized set of sounds. The result of this memory search task indicates the response instruction expected at S2; for instance, that a response has to be given either with the left or with the right hand, depending on the result of the search.
Referring back to FIG. 2, within the first second after S1 a slow wave complex can be seen which consists of a positive (slow wave) deflection (the “PSW”).
Later in the S1-S2 interval, a slow negative shift develops, which reaches a negative maximum immediately before S2. This shift is called the contingent negative variation (CNV).
When the S1-S2 interval is sufficiently long, three seconds or more, the negative shift clearly consists of two parts. The first part, called negative slow wave (NSW), is maximal at the frontal positions, between about 0.5 and 1 second after S1.
PSW has a parietal dominance and is assumed to reflect the outcome of stimulus evaluation and has been found to attenuate when the task is more difficult.
NSW has a frontal maximum, and some authors have found that it is lateralized in the right hemisphere. It is often regarded as part of an orientation reaction, because it is affected by the physical characteristics of S1, such as intensity, probability, and modality. The NSW is larger when the task at S1 is more difficult.
The CNV is mainly related to motor response preparation; its amplitude depends heavily on the task demands at S2, and is affected by task variables as speed/accuracy instructions, and the duration of the S1-S2 interval. CNV has the largest amplitude at the central electrodes.
CNV and Readiness Potential (BP)
CNV and BP are often considered to reflect the same process, since they are maximal at the same positions on the scalp, and immediately before a response is given. The difference between CNV and BP is, however, that the first is derived as a stimulus-locked potential, whereas the latter is derived relative to the response.
Whereas the BP is specific to motor readiness, and is concentrated over the primary motor cortex, the CNV is associated with more cognitive aspects of anticipation, and is generally localized to frontal and frontal-central cortex.
Topographic Plots of Bereitschafts Amplitudes with Different Types of Movement
In order to get a good picture of overall Motor Related Potential distribution effects, including lateralization effects, experimenters employ a two dimensional interpolation scheme from data collected through a multiple electrode arrangements. This technique facilitates the visualization of the topographic distribution of BP amplitudes over the entire scalp surface.
(Briefly, the data processing method consists of: first, the amplitude of the total negativity is measured algorithmically for each electrode position of the whole scalp arrangement. Because the data includes a large amount of high-frequency noise, a mean of 20-50 ms of voltage data is used to estimate the potential at the start and end of each BP waveform. The end value is then subtracted from the start value, thus yielding (in most cases) a positive magnitude for the negativity. These data is then combined with the known relative coordinates of each electrode to generate a two-dimensional grid interpolation of the overall negativity values. In addition, before the interpolation is applied, the EOG and EMG electrodes are removed from the data set as their given coordinates are figurative and they generally showed no evidence of significant waveforms.)
Free Movement
If subjects are asked to initiate a voluntary extension of the middle and ring fingers necessary and sufficient for production of a reliable Extensor Communis EMG signal of the right hand only, the recorded MAC is as in FIG. 3 (left) and the topographic distribution of the signal following the data processing described above is as in FIG. 3 (right).
CP2 (central parietal 2, slightly to the right of the central line and just below towards the back of the brain, relative to the point in FIG. 1). In the graph in FIG. 3 left shows the activity contra lateral to the movement. The translation of color into gray scale is as follows: Colors are gradual changing values. The two rounded dark areas in the center are positive peaks and the other dark areas are negative peaks.
In this case the voluntary movements are completely at the will of the participant, although a rough guideline is given to the subject to leave at least two to three seconds between movements. Accordingly, we can see most of the features described previously.
Synchronized Movement
If instead of moving the finger at will the subject is to initiate movement according to a self-maintained, even, metrical pulse with a rough guideline frequency of around 0.5 Hz, the resulting potentials are different. Ideally this will produce MAC events at the steady, regular rate of 0.5 Hz. These events should be phase locked to the subject's internal pulse. This is a ‘synchronize’ condition.
The resulting topographical maps are shown in FIG. 4 where BP amplitude is visualized as a color spectrum mapping (shown as grey scale). In these plots, a clear distribution effect of the experimental condition is even more evident than in the previous case of free movement. The dark area on the bottom is positive values and the dark area on the top is negative values.
As before, the movement consisted in the extension of the middle and ring fingers of the right hand sufficient for production of a reliable Extensor Communis EMG signal. It is clear that in this condition, the spread is mostly towards the left side of the head.
Applying the topographic algorithm described in the previous section, (mean of 20-50 ms of voltage data at beginning and end each signal), it can be seen that there are two spots of maximal amplitude of the whole signal: one is central and the other is slightly to the side contra lateral to the movement.
Syncopate Movement
In this case, the subject is again instructed to maintain an internal metrical pulse. However, in this trial the subject is instructed to initiate finger movements exactly counter to the pulse. That is, the movements should be phase shifted by half a period from the maintained internal pulse. Ideally this will produce MAC events at a frequency of 0.5 Hz, phase-shifted by 1 second from the internal pulse. This is the ‘syncopate’ condition. Results are shown in FIG. 5. (high values are at the bottom of the image, low on top).
From FIGS. 3-5 it can be seen that a clear distribution of the experimental condition is evident. First, there is an apparent spreading in the location of maximum BP amplitude in all the experimental conditions relative to the “Free Movement” condition. In the “Synchronized” conditions, the spread is mostly towards the left side of the head. In the “Syncopate” condition some spread to the right side is also present.
One potential effect is the appearance of CNV in the syncopate conditions. As described in a previous section, CNV usually marks expectation or anticipation in non-motor regimes.
Motor Imagery as Activator of Cortical Activation
Very recently a technique named “mirror therapy” has been reported to be used to activate unused cortical networks and help to reduce the pain associated with cortical abnormalities following injury as occurs in phantom pain and stroke. Briefly, “mirror therapy” involves the movement of a limb inside a mirror box such that visual feedback of the affected hand is replaced by that of the (reflected) unaffected hand. There is therefore an attempt to reconcile motor output and sensory feedback and to activate pre-motor cortices. In his last article, Moseley 2004 writes: “The mechanism of the healing effect of this technique, although not clear, may involve the sequential activation of cortical pre-motor and motor networks, or sustained and focused attention to the affected limb, or both.”
Slow Cortical Potential (SCP)
SCP in general has been extensively used by Prof. Neils Birbaumer and his group in the University of Tuebingen, Germany. SCPs are potential shifts in the scalp-recorded EEG that occur over 0.5-10 sec. Negative and positive SCPs are typically associated with functions that involve cortical activation and deactivation, respectively. Healthy subjects and neurological patients can attain reliable control over their SCPs amplitude and polarity recorded at frontal and parietal locations by means of sensory (e.g., visual and audio) feedback. In addition, subjects can learn to control SCP differences between the left and right hemisphere.
Mu Rhythm
When a person is at rest, his sensorimotor cortex generates an 8-13 Hz electro-encephalographic rhythm referred to as the “Rolandic mu rhythm”. As soon as the person starts to execute a movement and the motor cortex is activated, the mu rhythm is attenuated or disappears. The mu rhythm is present in most, if not all, adults, it is generated by a thalamo-cortical network and it is strongest when no active processing is occurring. Decreases in mu amplitude possibly indicate that the underlying cell assemblies have become desynchronized (hence, lower mu rhythm amplitudes are recorded over the scalp). The mu rhythm also desynchronizes when subjects are only observing (but not executing) a movement and the degree of desynchronization reflects the level of active processing; for instance, it is greater when a person is performing a precision grip than during the performance of a simple hand extension. These rhythms also respond to imagination of movement in a pattern similar to that during the planning of a movement. For instance, subjects with limb amputation that mentally mobilize the missing limb, show a blocking effect of mu rhythm while imagining the movement.
Feedback training leads to increase in mu desynchronization; and the ability of subjects to manipulate the sensorimotor mu rhythm has been recently used by Wolpaw et al. group to act as a brain-computer interface based on a binary signal. For instance, subjects can learn to produce similar or differential mu activity over the two hemispheres in order to control left or right movement in a three-dimensional video game.
Sometimes differences of amplitude in both hemispheres are recorded for two frequency band rhythms to allow for a bi-dimensional movement over the screen. For instance, in Wolpaw et al (Jonathan R. Wolpaw and Dennis J McFarland 2004. Control of a two-dimensional movement signal by a noninvasive brain-computer interface in humans. PNAS 2004, vol. 10, no. 51: 17849-17854), the disclosure of which is incorporated herein by reference, use vertical movement of a cursor was controlled by a 24-Hz beta rhythm and horizontal movement by a 12-Hz mu rhythm recorded at left- and right-side scalp electrodes locations C3 and C4 over the sensorimotor region. Vertical correlation is greater on the left side, whereas horizontal correlation is greater on the right side.
FIG. 9 shows on the left the various positions on the screen towards which the user learns to move the cursor from a center position. On center/right of FIG. 9 there are the recordings of brain potential when a user tries to move to Target 1 (up), Target 6 (down), Target 3 (right) and Target 8 (left). It can be seen, for instance, that in order to move to Target 1 (up) the user needs to increase Beta rhythm (24 Hz) recorded over the C3 and in order to move to Target 8 (left) the user needs to reduce the mu rhythm (12 Hz) recorded over the C4 region. In the figure, vertical control has two positive peaks as shown, while in horizontal control, the left peak is a negative peak and the right peak is a positive peak.
Many studies have demonstrated that humans can learn to control μ-rhythm amplitudes independent of actual movement and use that control to move a cursor to targets on a computer screen (Walpow et al., 1991; McFarland et al., 1993; Pfurtscheller et al., 1993), or to control an artificial hand attached to the paretic hand (Pfurtscheller et al., 2003). None of these studies, apparently was conducted in stroke patients or other patients with brain damage, such as that due to traumatic brain injury.
Cortical Reorganization after Stroke
Return of voluntary arm movements is one of the most important goals during stroke rehabilitation to avoid long-term disability in activities of daily living (ADL) function.
Some studies have demonstrated recruitment of areas adjacent to the brain lesion or ipsilateral motor regions of the unaffected hemisphere after complete recovery from upper extremity motor impairment. Rossini et al. (1998) for example, who used brain imaging methods as functional magnetic resonance imaging (fMRI), transcranial magnetic stimulation (TMS), and magnetoencephalography (MEG) to examine a patient who fully recovered after stroke, found an asymmetrical enlargement and posterior shift of the sensorimotor areas localized in the affected hemisphere with all three techniques.
Nelles et al. (1999), used serial positron emission tomography (PET) to study the evolution of functional brain activity within 12 weeks after a first subcortical stroke. Six hemiplegic stroke patients were scanned twice (PET 1 and PET 2). At PET 1, activation was observed in the bilateral inferior parietal cortex, contralateral sensorimotor cortex, and ipsilateral dorsolateral prefrontal cortex, supplementary motor area, and cingulate cortex. At PET 2, significant increases of regional cerebral blood flow were found in the contralateral sensorimotor cortex and bilateral inferior parietal cortex. A region that was activated at PET 2 only was found in the ipsilateral premotor area. Based of their findings, Nelles et al conclude that recovery from hemiplegia is accompanied by changes of brain activation in sensory and motor systems, and that these alterations of cerebral activity may be critical for the restoration of motor function.
Johansen-Berg et al. (2002) examined seven stroke patients with fMRI twice before and twice after a home-based two weeks rehabilitative therapy. They found that therapy-related improvements in hand function correlated with increases in fMRI activity in the premotor cortex and secondary somatosensory cortex contralateral to the affected hand, and in superior posterior regions of the cerebellar hemispheres bilaterally. As the former studies, these results suggest that activity changes in sensorimotor regions are associated with successful motor rehabilitation.
In addition, accumulating evidence in stroke patients suggests that rehabilitation techniques with repetitive training of functional movements have significant effects on recovery of motor skills and cortical reorganization. Lipert et al. (2000), for example, evaluated reorganization in the motor cortex of stroke patients that was induced with an efficacious rehabilitation treatment. Before treatment, the cortical representation area of the affected hand muscle was significantly smaller than the contralateral side. After a 12-day-period of constraint-induced movement therapy, the muscle output area size in the affected hemisphere was significantly enlarged, corresponding to a greatly improved motor performance of the paretic limb. Shifts of the center of the output map in the affected hemisphere suggested the recruitment of adjacent brain areas. In follow-up examinations up to 6 months after treatment, motor performance remained at a high level, whereas the cortical area sizes in the 2 hemispheres became almost identical, representing a return of the balance of excitability between the 2 hemispheres toward a normal condition.
Luft et al. (2004) tested whether specific rehabilitation therapy that improves arm function in stroke patients is associated with reorganization of cortical networks. Patients were randomly assigned to bilateral arm training (n=9) or standardized dose-matched therapeutic exercises (n=12). Both were conducted for 1 hour, 3 times a week, for 6 weeks. Within 2 weeks before and after the intervention, brain activation during elbow movement was assessed by fMRI and functional outcome was assessed using arm function scores. Patients in the first group (bilateral arm training) but not in the second group increased hemispheric activation during paretic arm movement. Significant increased activation was observed in the contralesional cerebrum and ipsilesional cerebellum. These findings suggest that bilateral arm treatment induces reorganization in contralesional motor networks and provide biological plausibility for repetitive bilateral training as a potential therapy for upper extremity rehabilitation in hemiparetic stroke.
Summary
Numerous studies have demonstrated that the damaged brain is able to reorganize to compensate for motor deficits. Rather than a complete substitution of function, the main mechanism underlying recovery of motor abilities involves enhanced activity in preexisting networks, including the disconnected motor cortex in subcortical stroke and the infarct rim after cortical stroke. Involvement of nonmotor and contralesional motor areas is consistently reported, with the emerging notion that the greater the involvement of the ipsilesional motor network, the better is the recovery. A better stroke recovery seems to take place if the changes in certain brain areas over time are such that the normal balance between the 2 hemispheres tends to reestablish. Thus, recovery is best when the brain regions that normally execute the function are reintegrated into the active network. Consistent with this view, intense rehabilitative procedures (both active and passive) have recently been shown to enhance activation of the ipsilesional motor area in parallel with improved motor function.
Consistent results were a dynamic reorganization that went along with recovery, over-activation of motor and non-motor areas in both hemispheres regardless of whether the task was active or passive leading to a decrease in the laterality index, and a return toward a more normal intensity while the affected hand regained function (Calautti and Baron, 2003). In some embodiments of the invention, as described below, such results are achieved using methods and apparatus as described herein.
Some components of brain activity (e.g., SCPs) can be controlled by increasing or decreasing the general brain activity, independent from the recorded site. Subjects can learn to control SCP differences between the left and right hemisphere (Rockstroh et al., 1990).
Brain-Computer Interface (BCI)
A BCI system measures particular components of features of EEG activity and uses the results as a control signal. Present-day BCIs determine the intent of the user from a variety of different electrophysiological signals. These signals include slow cortical potentials (SCP), P300 potentials, and μ (mu) or beta rhythms recorded from the scalp, and cortical neuronal activity recorded by implanted electrodes (referred as Brain-Machine Interface, BMI). They are translated in real-time into commands that operate a computer display or other device.
A BCI converts a signal such as an EEG rhythm or a neuronal firing rate from a reflection of brain function into the end product of that function: an output that, like output in conventional neuromuscular channels, accomplishes the person's intent. A BCI replaces nerves and muscles and the movements they produce with electrophysiological signals and the hardware and software that translate those signals into actions.
Many studies have demonstrated that healthy subjects and neurological patients can attain reliable control over their slow cortical potentials (SCPs) amplitude at vertex, frontal and parietal locations with operant learning. Moreover, subjects can learn to control SCP differences between the left and right hemisphere (Birbaumer et al., 1999; Birbaumer et al., 1988; Rockstroh et al., 1990). Successful learning using reinforcement and shaping of the response results in the acquisition of a new, non-motor skill (Birbaumer et al., 1999). Studies involving mu-rhythm (Walpow et al, 2002) confirm and extend these findings.
The following articles, some of which are referenced herein, have their disclosures incorporated herein by reference:
International Journal of Psychphysiology 9 (1990) 151-165 “Biofeedback-produced hemispheric asymetry of slow cortical potentials and its behavioural effects”, by B Rockstroh, T Elbert, N Birbaurmer and w Lutzenberger.
Wolpaw, J. R., McFarland, D. J., Neat, G. W. and Forneris, C. A. “An EEG-based brain-computer interface for cursor control” Electroenceph. clin. Neurophysiol., 1991, 78:252-259.
McFarland, D. J., Neat, G. W., Read, R. F. and Wolpaw, J. R. “An EEG-based method for graded cursor control” Psychobiology, 1993, 21: 77-81.
Pfurtscheller, G., Flotzinger, D. and Kalcher, J. “Brain-computer interface: a new communication device for handicapped persons” Journal of Microcomputer Applications, 1993, 16:293-299.
“Arm Training Induced Brain Plasticity in Stroke Studied with Serial Positron Emission Tomography” G. Nelles, W. Jentzen, M. Jueptner, S. Mu{umlaut over ( )}ller, and H. C. Diener. NeuroImage 13, 1146-1154 (2001) doi:10.1006/nimg.2001.0757.
“Control of a two-dimensional movement signal by a noninvasive brain-computer interface in humans” Jonathan R. Wolpaw and Dennis J. McFarland. PNAS Dec. 21, 2004 vol. 101 no. 51 17849-17854
“‘Thought’—control of functional electrical stimulation to restore hand grasp in a patient with tetraplegia” Gert Pfurtschellera, Gernot R. Mu{umlaut over ( )}ller, Jo{umlaut over ( )}rg Pfurtscheller, Hans Juürgen Gernerd, Ru{umlaut over ( )}diger Ruppd. Neuroscience Letters 351 (2003) 33-36
Nelles G, Spiekermann G, Jueptner M, Leonhardt G, Mu{umlaut over ( )}ller S, Gerhard H, Diener H C. “Evolution of functional reorganization in hemiplegic stroke: a serial positron emission tomographic activation study” Ann Neurol 1999; 46:901-909
“Hand motor cortical area reorganization in stroke: a study with fMRI, MEG and TCS maps” P. M. Rossini, C. Caltagirone, A. Castriota-Scanderbeg, P. Cicinelli, C. Del Gratta, M. Demartin, V. Pizzella, R. Traversal and G. L. Romanil. NeuroReport 9, 2141-2146 (1998)
“Correlation between motor improvements and altered fMRI activity after rehabilitative therapy” Heidi Johansen-Berg, Helen Dawes, Claire Guy, Stephen M. Smith, Derick T. Wade and Paul M. Matthews. Brain (2002), 125, 2731-2742
“Treatment-Induced Cortical Reorganization After Stroke in Humans” Joachim Liepert, MD; Heike Bauder, PhD; Wolfgang H. R. Miltner, PhD; Edward Taub, PhD; Cornelius Weiller, MD (Stroke. 2000; 31:1210-1216.)
Birbaumer, N. et al. (1999). “A spelling device for the paralysed.” Nature 398(297-298).
Birbaumer, N., et al. (1988). “Slow brain potentials, imagery and hemispheric differences.” International Journal of Neuroscience 39: 101-116.
Rockslroh, B., et al. (1990). “Biofeedback-produced hemispheric asymmetry of slow cortical potentials and its behavioral effects.” International Journal of Psychophysiology 9(2): 151-165.
Wolpaw, J. R., Birbaumer, N., McFarland, D. J., Pfurtscheller, G., Vaughan, T. M., 2002. Brain/computer interfaces for communication and control. Clinical Neurophysiology 113, 767-791.
Ward N S, Brown M M, Thompson A J, Frackowiak R S. Neural correlates of outcome after stroke: a cross-sectional fMRI study. Brain 2003; 126:1430-1448.
“Repetitive Bilateral Arm Training and Motor Cortex Activation in Chronic Stroke A Randomized Controlled Trial” JAMA. 2004; 292:1853-1861
“Functional Neuroimaging Studies of Motor Recovery After Stroke in Adults A Review” Cinzia Calautti, MD; Jean-Claude Baron, MD, FRCP, FMedSci (Stroke. 2003; 34:1553-1566.)